Credit-Risk Decision Process Using Neural Networks in Industrial Sectors

Abstract

Due to increased bankruptcies noted among companies (debtors) banks pay more attention to credit risk management. One of the most valid tasks in credit risk evaluation is the proper classification of potential good and bad customers. Reduction of the number of loans granted to companies of questionable credibility can significantly influence banks’ performance. An important element in credit risk assessment is a prior identification of factors which affect companies’ standing. The research focuses on determining which of the factors have the biggest impact on company’s solvency and which are redundant and therefore can be removed from future analysis. The other purpose of the research is to investigate and compare the results of two different structures of neural networks—the most common Multi-Layer Perceptron (MLP) and Radial Basis Function neural network (RBF). The conducted research bases on the financial reports of Polish companies in industrial sector and a credit risk analysis method applied in one of the banks operating on Polish market. The results of two different NN models are juxtaposed and compared with the real-world data. Moreover, the vulnerability analysis of entry data is carried out to find the most beneficial set of variables.